Digital twins: a myth buster

Digital twin
Murray Callander
September 8, 2021

There’s a lot of hype about digital twins in oil and gas. Much is well-founded, since the integration of the underlying technologies that make digital twins possible have potential to accelerate much needed digital transformation in the sector. But you can’t believe everything you read.

A digital twin is a digital representation of a real-world entity or system; it uses real-world data to create simulations that can describe how that entity (or a product or process within) is performing – or could perform.

We have been working with digital twins since their early incarnations in the late 2000s and have learnt a lot. In this Eigen Perspective we cut through the hype, bust some myths and present a more measured case for digital twin technology.

Digital twin technology relies on cloud computing which means the scope for simulating the real-world increases dramatically.

As more data, Internet of Things-enabled equipment and applications can be integrated and automated – to perform faster and more complex calculations, with increasingly sophisticated visualisation.
Digital twin is the realisation of the digital oilfield technology vision of the early 2000s, when Eigen first began working with bp on its first implementation in Azerbaijan.

It’s no exaggeration to say that in complex, physics-based industries like oil and gas, digital twin technology is probably the biggest technology game-changer since seismic imaging.

Few think it’s a short term craze: growth in the digital twin market is forecast across all sectors globally from $3bn in 2020 to $48bn in 2026, driven largely by the need to reduce costs (design, build and maintenance) and recover from, and respond to, the pandemic (Markets and Markets, 2020)

In complex, physicsbased industries like oil and gas, digital twin technology is probably the biggest technology game-changer since seismic imaging.

Myth 1: Digital twin is just visualisation and simulation; the rest is Emperor’s New Clothes.

Visualisation and simulation are key components of any digital twin, enabled and powered by giant leaps in cloud computing and capacity to operate near-real-time. But the game-changer in making digital twins a reality is knowledge graph technology.

At Eigen we use knowledge graph technology specialist Neo4j to build domain-specific data models, e.g. safety barrier data models, that act as the brain of the digital twin, calling on specific data or calculations to run specific actions e.g. simulations, visualisation, alerts or events.

Powered by the cloud and knowledge graph technology, the digital twin has emerged as one of the biggest technology breakthroughs in a generation, transforming industries as varied as pharmaceuticals, automotive and oil and gas. It’s far from Emperor’s New Clothes.

A knowledge graph is a model of a specific domain built by human experts using machine learning algorithms. It provides a structure and common interface for relevant domain-specific data and the relationships between those data. But it doesn’t need to house all data, it sits on top, linking together existing databases and datasets (even unstructured and real-time data) - with virtually infinite scale.

Myth 2: The digital twin is the Oracle, all-knowing, all powerful.

Digital twin technology is undoubtedly powerful, but like most technologies, it is limited by the quality and accuracy of the data it can access, how its various components have been designed – and how humans interact with them.

We wouldn’t characterise the satnav as all-knowing because it can access a data model (or knowledge graph) to optimise our route from A to B. The same is true for digital twin technology: the user sets the objective, defines any parameters and conditions, and ultimately chooses whether to accept or ignore the advice.

A good example of this is where Eigen worked with Lundin to build a digital twin of its blowdown system, using the Ingenuity platform. Blowdown valves on plant are designed to depressure equipment fast, to make it safe.

Blowdown performance must meet industry standards and is subject to safety audit.
Without the digital twin, managing blowdowns was a time-consuming manual job, requiring data gathering from multiple documents and systems, exporting to Excel for analysis and pasting data and charts into Word-based reports.

With the blowdown digital twin the Lundin operations team are now only alerted when a blowdown occurs that does not meet agreed performance criteria. The team can access a live, online report of all previous blowdown performance providing full history for analysis and audit.

While the digital twin is not all-knowing, it provides assurance, and frees up high cost engineering time for more value adding work.

Digital twin technology provides an aid to humans, who ultimately retain power to act.

Myth 3: You only capture value from a digital twin that covers your entire asset.

In many ways, this kind of monolithic thinking has halted progress of digital transformation: the comforting, but ultimately false, logic that housing all data and systems in one place is the only route to control.

Systems, like those for oil and gas operations, are simply too big, too complex and too dynamic to be designed and operated as a single digital twin. 

Moreover, value would haemorrhage from any such project, owing to time lost in design, as well as overlooked low hanging fruit.

At Eigen we work with clients to solve a specific problem or use-case, using technology to bring information and insight to the client to accelerate decision-making and drive team productivity. Then we work incrementally, use-case by use-case to extend the capabilities of the digital twin.

For instance, we built a maintenance backlog digital twin for one North Sea operator, enabling them for the first time to calculate and visualise barrier risk and use this to help prioritise and plan maintenance. Using the digital twin, the operator reduced the maintenance backlog on Safety & Environmentally Critical Equipment by 90% in three weeks

90% Reduction in maintenance backlog in three weeks for one North Sea operator using Eigen’s digital twin technology

Myth 4: There’s a digital twin one-stop-shop who can meet all your needs.

Unfortunately there’s not, despite what some vendors may imply. At least not for the physics-based world of oil and gas. Scale and complexity prohibit such a thing, but perhaps of more significance is the specialist know-how layered into every piece of equipment within oil and gas systems, which no generalist vendor could replicate. 

For example, the sophistication of a digital twin built by the original equipment manufacturer of a compressor would far exceed any digital twin a generalist could design. And why would it, when all it needs to know is a limited subset of the data the compressor produces when it shuts down, needs attention or starts up? 

Digital twins therefore are the aggregation and integration of multiple digital twins, with the least redundancy and duplication to perform fast and accurately. Eigen for instance works with multiple specialist vendors to take specific data they produce to feed its digital twins. With bp, Eigen’s Ingenuity platform takes work order data from Maximo, to feed its Safety Barrier Health digital twin. Similarly, we integrate with OSISoft’s PI historian to point to operators’ temperature, pressure and flow rate data. And we access data from plant and equipment direct from ABB and Honeywell control systems. 

There’s simply no value for the client in copying data into the digital twin, when it can be linked and its integrity managed by an expert. Rather than a digital twin vendor, clients tell us they see Eigen as the ‘glue’ that integrates their digital twins to drive efficiency, productivity and improved levels of safety

There’s simply no value for the client in copying data into the digital twin, when it can be linkedand its integrity managed by an expert

Myth 5: Digital twins only tell you about the present and the future.

Digital twin technologies can certainly help inform users what’s going on right now – and through simulation and visualisation can help predict the future. But they increasingly enable us to look back and retain “project memory”. 

We are now just beginning to see the first oil and gas projects come online that have been fully designed as a digital twin, through concept development, design and build – and are now moving into operation equipped with digital twin technologies. 

This means for the first time, we are also able to use digital twin technology to look back to prior decisions, previous designs, engineering rationales etc; the digital twin becomes the repository of project memory. 

We have been working with one major operator on its latest project in the Gulf of Mexico, to build a digital twin of its safety barriers – this will enable us to look backwards, as well as in real and future time, harnessing the power of project memory as well as predictive capability. We have also built something similar for the new Edvard Grieg tiebacks Solveig and Rolvsnes, operated by Lundin Energy.

Myth 6: Digital twins rely on advanced AI.

There is undoubtedly a role for artificial intelligence in equipping digital twins with even more advanced advisory capabilities, particularly making recommendation to users. 

By interrogating huge volumes of safety, operations, subsurface, production and project data AI could doubtless improve early warnings or identify new opportunities that are challenging for a human. 

Yet, today’s digital twins, equipped with today’s technologies (Neo4j, Elastic, Docker, Python scripts and Eigen’s Ingenuity platform), are able to integrate massive volumes of data from multiple sources, and systems and applications in real-time, with sophisticated visualisation – and interfaces designed around the user (e.g. Search, Mobile, Screen Presenter, Interactive Dashboard). 

These integrate to deliver low hanging fruit, like converting a document manually prepared each week of copy/ pasted data and charts into an online, mostly automated report for Lundin. 

And also giant leaps in digital transformation, like building one of the most sophisticated safety barriers digital twins in the industry, used by Vår, Lundin and bp.

Myth 7: Digital twin technology is only an option for the big players.

Digital twin technology is not cheap nor is it a one-off investment. But because it can often deliver a quick return and payback, it is an accessible technology option for all operators, regardless of size. 

In many ways digital twin technology is the digital transformation, and as such has to be evaluated as an investment to deliver future value from cost savings, efficiencies, accelerated production – and potentially new resources. 

Many operators we work with at Eigen begin fearing a wholesale change of systems and processes, and while this can happen over time – and indeed may be necessary – we favour more incremental implementation, solving one use-case at a time, demonstrating value along the transformation journey. 

As the Markets and Markets digital twin report (2020) identified, cost reduction is a major incentive for investment in digital twin technology. When the true cost of many manual jobs is calculated, investment in digital twin technology, particularly problem-solving use-case by use-case, often has a compelling return and quick payback.

Eigen Perspectives are straightforward, accessible articles that share what we have learned over fifteen years working as digital problem-solvers in oil and gas.

For more information on Digital Twin technologies and how Eigen can help, contact

Written by
Murray Callander
Posted on
September 8, 2021
September 20, 2021

Digital frontline

September 8, 2021

Digital Transformers

September 6, 2021

Digital twin: technology of now

June 30, 2021

To Copy Or Not To Copy – That Is The Question

May 30, 2021

How to make a Digital Twin of a Bowtie – Part 2

April 27, 2021

How to make a Digital Twin of a Bowtie – Part 1

You may also like


Distributed Component Object Model (DCOM) is a Microsoft remote communication protocol at the heart of many real time data links.

Knowledge Graphs

Context and clarity through dynamic monitoring and visualisation

Knowledge Graphs

Glowing net and spirals
There is definitely something intuitive about building knowledge graphs, after all, they mirror the way we as humans acquire knowledge and retain it through connections and context.

Digital Transformation

The trend in many mature basins, just like the UK and Norwegian continental shelves, has been one of steady divestment by the supermajors and a new generation of smaller oil and gas operators.

Electric cars

The running cost of an electric car depends on where you charge it and can vary from 3p to over 35p per mile.
In short, electricity prices vary hugely compared to petrol or diesel; you can pay anywhere from 5p/kWh to 70p/kWh so it costs me anywhere from 3p/mile to 35p/mile depending on where I charge.

Electric cars

The range of an electric car is less important than how quickly you can recharge it
Would I recommend an electric car? Yes, definitely. They are nicer to drive, much cheaper to run and need less maintenance. Range isn’t the issue, maybe the real question is “are you prepared to change your habits?”.

Electric cars

reliable destination charging is key to making electric car ownership work for most people.
When you say you own an electric car everyone’s first question is about that magic word “Range” – “How far can it go on a charge?”. Having owned an electric car for over six years now I can tell you that range is much less important than the ability to recharge at your destination. Let me explain why.

Safety Barriers

At Eigen, we build and deploy live data models as part of a digital twin for oil and gas facilities. One of the challenges in doing this is how to combine the physical and the abstract into a working data model. And a great example of this is the bow tie – a diagram showing all the protective measures against hazardous events and their potential consequences.

Knowledge Graphs

The automobile and industrial designer Freeman Thomas said “We were promised a simpler life, and technology has only complicated our lives.” In complex domains, like oil and gas, technology may well have simplified some things but it has certainly complicated others, particularly where data is concerned.

Decision making

Death and taxes may historically have been the only two certainties in life, but we might consider a third – increasing volumes of data. With ubiquitous sensing, unlimited cloud storage and insatiable demands from every walk of life, data is growing exponentially.

Digital twin

Much of the focus of digital transformation is given to big data, clever systems and game-changing innovations, like robotics, AI and IoT. But what is often missing is the focus on people, who have been irrevocably transformed by digital.

Digital twin

We formed Eigen almost fifteen years ago, although the original founding team had all been working in oil and gas for the previous decade. As individuals, then later as Eigen, we worked with bp to deploy its breakthrough Field of the Future programme in Azerbaijan, and have since been part of many oil and gas operators’ digital transformation journeys.

Digital twin

The concept of digital twin has a futuristic quality at a time when robotics and artificial intelligence also are coming of age. But whilst there’s no disputing the transformative impact of digital twins, in industries as wide-ranging as construction, marine, agriculture and oil and gas, what makes up a digital twin is technology of now.

Decision making

One of the problems of asset-heavy industries, including Oil & Gas and Petrochemical operations, is decision-making for maximum impact, under resource constraint conditions. These industries typically manage an array of highly interconnected assets and relationships which bear on the impact of the decision.

Digital twin

When it comes to digital transformation and learning how to discover insights hidden in data, one question that regularly arises is whether we should copy all of our data into a data warehouse or data lake, or alternatively connect directly to the source systems.

Digital twin

In this blog series we’re looking at how you turn a bowtie diagram into an online Digital Twin showing live barrier health information. The previous part looked at the level of detail required to make a working data model in practice. In this instalment we’re looking at the challenge of assessing the health of protective functions not directly related to the physical SECEs (Safety and Environmentally Critical Equipment).

Digital twin

In this blog series we’re looking at how you turn a bowtie diagram into an online Digital Twin showing live barrier health information. In this instalment we’re considering the level of detail required to make a working data model in practice.

Digital twin

How do you turn a bowtie diagram into an online Digital Twin showing live barrier health information?

Complex Subsea Hydraulic System

Imagine the situation: Hydraulic fluid pressure is dropping fast; supply tank levels are also dropping. You are pretty sure there is a big leak in the system, but you have no idea where it is. Pretty soon the operation will be compromised: The pressure will be below the minimum required and you may be forced to shut down the field. It is a scary picture.

What makes a good datamodel?

Wireframe rendering of turbojet engine and mirrored physical body on black background. Digital twin concept. 3D rendering image.
There is a lot of buzz nowadays about “Digital Twins”, and every Digitalisation Director out there is thinking they need to have a Digital Twin of their assets. It makes sense: If you have a digital representation of your assets; you can simulate, test scenarios, plan operations and visualise data within the context afforded by the twin, without having to visit or “be present” in the real asset, especially if these assets are remote.


Agile development software business web computer agility nimble quick fast start up concept
Ah, the old ones still the best ????. But such questions can sometimes be worth considering. For example, when is Agile not Agile?

Open Source

Two months ago I switched my phone from an Apple iPhone 7 to an Android based POCO X3 NFC. At the same time one of my daughters upgraded their phone to an iPhone XR so we’ve been comparing experiences. So what’s it been like??


2021 futuristic tehcnology trend concept, hand man pointing in concept futuristic in 2021 coming year the year of artificial intelligence ,big data, iot, augmented reality, machine learning
As the COVID-19 pandemic hit us round about March 2020 and the Oil prices crashed, all of us working for the Oil & Gas industry panicked and prepared for the worse. In previous market downturns, budgets had been slashed and projects cancelled, but this time something quite different happened, at least in the Digitalisation space.


What has Agile ever done for us?
In an earlier blog, I wrote about what Agile is. And that’s all well and good, but what has Agile ever done for us?

Open Source

Using Open Source to avoid lock-in to expensive proprietary technologies
I’ve been using Apple products since 2007. Every smart phone I’ve had has been an iPhone but yesterday I ordered a new Android phone. I’m so fed up with shelling out around £1000 every couple of years but I’ve been trapped because all my music is in iTunes, all my family have iPhones and we run Macs*. Man, I hate being locked in! I wouldn’t mind if it was still value for money, but it’s not – I feel like I’m being taken for a ride now.


In November (this month as I write this) we have not long closed out possible two of the best custom software projects we have done. Sure there were bugs in the testing and there will be more found by the users I’m sure, but what made them so good? We hit the schedule to within a week on both of them and there were no scope changes over the entire project, so we hit the budget as well. This is particularly impressive for one of them because it was a completely new piece of software, written from scratch; taken from powerpoint to in production in 5 months.


percent sign on screen. interest rate
This picture makes me laugh and if it makes you laugh too then you’ve probably been involved in purchasing or project delivery at some point.

Is your Oil field being exploited at its best?

A man holding a flask of oil. Oil rigs in the background. Oil production in Russia. Sunset. Oil pumping
Production Well Testing is an essential part of managing an Oil & Gas field if the field has more than one well and production measurement resources are limited, as is generally the case.


Agile Software Development Business Internet Techology Concept
You have probably heard software developers talk about being Agile, and maybe wondered what they mean. Perhaps they do star jumps and touch their toes before breakfast, or write code while limboing under desks? Well, it’s nothing like that… read on to find out more!


Eigen and Wintershall Dea partnership
I’m really excited about the long term partnership we have recently signed with Wintershall Dea for a number of reasons not least because it’s a great team and we have a lot of fun working together!

Decision Bridge Technology

Back in July I asked myself the question “Why are we still in business after 11 years?”. A small specialist company in a big competitive industry. We’ve been through 3 downturns now and we continue to develop longstanding relationships with some of the most innovative operators in the business. Clearly we are doing something right but what is it?

Decision Bridge Technology

Young musician playing acoustic guitar and singing,
Did you ever learn to play “Sweet child of mine” on the guitar? Like almost every other teenager who had a guitar I learned the opening riff to “Sweet child of mine” and I remember playing it in a guitar shop when I was testing out an amp, only to have one of the staff point to a sign on the wall that said anyone playing “sweet child of mine” would be thrown out the shop and told never to return!

Is your Oil field being exploited at its best?

Smart city and abstract dot point connect with gradient line and aesthetic Intricate wave line design , big data connection technology concept .

In our previous blog, the discussion was centred on assessing how well an asset is exploited, as a key concern for the field manager and the asset team. The Integrated Choke Model (ICM) was proposed as a model to identify opportunities for improving the free cash flow as the fundamental objective during periods when the oil price is low.


Businessman holding tablet and showing holographic graphs and stock market statistics gain profits. Concept of growth planning and business strategy. Display of good economy form digital screen.
The world was shocked to learn about the massive explosion in the port of Beirut: This has been a terrible tragedy with an astronomical cost in terms of lives lost and damage to property; one could even argue, to the whole country, as even the strategic grain reserves were hit.

Is your Oil field being exploited at its best?

Is your Oil field being exploited at its best?
A typical question for a Petroleum Field Manager is how well the asset is being exploited and if this exploitation is efficient and generates enough value for Operator and Partners. This article explores some answers based on our understanding of the challenges that appear throughout the life of the field, and to how to be prepared to transform them into opportunities.

What makes a good datamodel?

What makes a good datamodel?
Now we are potentially getting right in amongst the weeds and many a data model gets stuck in the weeds because this issue is not understood.

What makes a good datamodel?

What makes a good data model? | Eigen
Once upon a time, in a land far far away, hard disks and memory were small and really expensive. In 1969 the computer that landed man on the moon only had 4k of RAM and 72k of ‘disk’ (ROM) storage.

What makes a good datamodel?

What makes a good data model? | Eigen
Having the best datamodel in the world won’t help you grow your digital capabilities if the only way to access it is through proprietary drivers.

What makes a good datamodel?

What makes a good data model? | Eigen
Google search is not a validated datamodel. Nor is it comprehensive, but it’s good enough for most of us.